Closed WizardOfMenlo closed 4 years ago
Something interesting is that criterion automagically generates an html report of the benchmarks, so if you look in target/criterion you can find some quite cool visualizations
Thanks, this looks good. Now has a conflict however due to mergint the other PR.
Oh yeah that is just the print that I commented since it was skewing benchmarks, I also made a small change that uses random permutations for hopefully more representative results
So, this is a simple example of what benchmarking could look like. You can run it with cargo bench. I have just benched (a b)^-1 vs b^-1 a^-1 on a couple of permutations that I came up with (so it is not really representative, I think a random permutation will help a lot for benches).